Tool-Augmented Agent for Closed-loop Optimization,Simulation,and Modeling Orchestration
Quick Answer
COSMO-Agent is a tool-augmented RL framework that optimizes CAD-CAE processes by enabling LLMs to manage geometric revisions under constraints.
Quick Take
COSMO-Agent is a tool-augmented RL framework that optimizes CAD-CAE processes by enabling LLMs to manage geometric revisions under constraints. It significantly enhances small open-source LLMs' performance in constraint-driven design, outperforming larger models in feasibility and stability. An industry-aligned dataset supports realistic training across 25 component categories.
Key Points
- COSMO-Agent addresses the CAD-CAE semantic gap through interactive reinforcement learning.
- A multi-constraint reward system ensures feasibility and toolchain robustness.
- The framework outperforms large open-source and strong closed-source models in design tasks.
- An industry-aligned dataset includes 25 component categories for effective training.
- Experiments show substantial improvements in efficiency and stability for small LLMs.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Iterative industrial design-simulation optimization is bottlenecked by the CAD-CAE semantic gap: translating simulation feedback into valid geometric edits under diverse, coupled constraints. To fill this gap, we propose COSMO-Agent (Closed-loop Optimization, Simulation, and Modeling Orchestration), a tool-augmented reinforcement learning (RL) framework that teaches LLMs to complete the closed-loop CAD-CAE process. Specifically, we cast CAD generation, CAE solving, result parsing, and geometry revision as an interactive RL environment, where an LLM learns to orchestrate external tools and revise parametric geometries until constraints are satisfied. To make this learning stable and industrially usable, we design a multi-constraint reward that jointly encourages feasibility, toolchain robustness, and structured output validity. In addition, we contribute an industry-aligned dataset that covers 25 component categories with executable CAD-CAE tasks to support realistic training and evaluation. Experiments show that COSMO-Agent training substantially improves small open-source LLMs for constraint-driven design, exceeding large open-source and strong closed-source models in feasibility, efficiency, and stability.
| Comments: | 8pages,3figures |
| Subjects: | Artificial Intelligence (cs.AI); Graphics (cs.GR) |
| Cite as: | arXiv:2605.20190 [cs.AI] |
| (or arXiv:2605.20190v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2605.20190 arXiv-issued DOI via DataCite |
Submission history
From: Liyuan Deng [view email]
[v1]
Wed, 1 Apr 2026 14:14:09 UTC (1,981 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →Procedural Memory Distillation: Online Reflection for Self-Improving Language Models
Procedural Memory Distillation (PMD) enhances reinforcement learning by converting cross-episode signals into reusable memory, improving Qwen3-8B and OLMo3-Instruct-7B models by 3.8-5.5% on SCIKNOWEVAL and 7.9-13.6% on . The co-evolution of policy and memory allows for more effective self-supervision, demonstrating significant performance gains when both components are active.